1. Field of the Invention
The present invention relates generally to add-on electronic devices for providing additional or new operating features for existing electronic systems, and more particularly to a method and apparatus for cooling components in such add-on devices. The invention further relates to methods of dissipating heat in plug-in type add-on circuits having at least one microprocessor coupled to other data processing elements.
2. Related Technology
In recent years, digital electronic equipment, such as, personal computers, word processors, work stations, and other electronic equipment using built-in microprocessors, such as printers, facsimile machines, memo devices, musical instruments, cooking equipment, and cameras, has found extensive use throughout large segments of society. In addition, other widely used equipment such as automobiles, robots, numerically controlled machines, and a variety of other electrified products, now make use of microprocessor technology.
The application of programmable digital logic to equipment operation makes more flexible control possible compared to that obtained with simple feedback controls previously used with various fixed hardware designs. In addition, using programmable logic, operating functions are essentially altered by simply changing command software. One advantage of this approach is that totally different control operations are obtainable for a given piece of equipment or hardware by simply modifying the content of program storage or memory elements, such as ROMs, that store specific processing or program steps. Moreover, smaller incremental changes in function, such as design revisions, can be advantageously implemented by only upgrading software.
However, the ultimate capabilities of processor controlled electronic equipment are determined by the capabilities of the processor itself. That is, each processor is itself finally limited by operating characteristics such as the maximum number of processing steps obtainable per unit time, the maximum number of data bits that can be processed at one time, the width of any data or command transfer buses, and so forth. As a result of these limitations, achieving improvements by merely upgrading software versions is at best limited to improving equipment ease of use. In reality, it has not been possible to achieve large improvements in operating features for existing electronic equipment.
At the same time, improving or upgrading software versions often requires replacing a ROM or other memory element containing the software. It is much more difficult to obtain access to or change software when replacement of such code containing ROMs is required. As a result, revising software to improve equipment operation is actually very difficult unless the particular piece of electronic equipment is already scheduled for a ROM exchange, different ROM version, at the time of its initial design, or unless the software can be supplied on a replaceable medium such as a flexible disk and used to modify stored program material.
For some applications, devices called "accelerators" are used in an attempt to improve overall function or capabilities by completely replacing key control components, such as microprocessors, within electronic equipment which otherwise limit operation. This type of hardware "upgrade" is commonly encountered with personal computers. However, this approach requires replacing a microprocessor which is generally located on a "motherboard" within the apparatus and represents a task that is beyond the skill of most equipment users. Furthermore, for typical consumer electronic equipment such as the previously mentioned printers, facsimile machines, musical instruments, cooking equipment, cameras, automobiles, etc., absolutely no consideration is commonly given to providing for such improvements or upgrading functionality and no such hardware option exists. A good example of this lack of planning is seen in relation to page printers which are manufactured for use with computers.
In recent years, page printers, such as laser printers, have enjoyed widespread distribution and are rapidly becoming the leading device for high-speed data and image output from computers. The resolution of laser printers typically ranges from 240 to 800 dots per inch (dpi), and printing speed is on the order of several pages a minute. Such printers principally employ an electrophotographic printer element, such as a xerography unit, which uses a photo-sensitive drum as part of the printing engine. After the printer has received and stored one page of image data (or blank), image processing steps, that is, electrostatic charge, exposure, toner application and image transfer, take place continuously in synchronization with rotation of the photo-sensitive drum.
Therefore, the page printer memory capacity for image development or processing must be sufficient to store at least one page of image data at a time. If no image data compression is employed, this capacity is determined by the printer resolution being used and the page size to be accommodated. For example, if a resolution of 300 dpi and a page size of 8 by 10 inches are considered, the printer may handle as much as 8.times.300.times.10.times.300 or 7,200,000 dots, or pixels, of image data, which requires at least 0.9 megabytes of memory capacity. Moreover, if the print or image input data is in the form of a bit mapped image to be printed, the printer only needs to accept and sequentially store this data before image processing. The processing speed for this type of operation generally depends on, and is limited by, the data transfer rate. Since parallel data transfer, such as that complying with the Centronics specification standard, occurs at a very high rate, it is unlikely that data transfer of bit images will occur slower than the printing capability of the xerographic unit.
However, where printers receive and process other types of data, such as character codes, line positions, and line and character pitch, and then develop this data into a page image, or receive programs that describe the page using a page description language (PDL) and then interpret and process this information to generate a page image, requires arithmetic processing and generation of bit mapped images from the input print data. In comparison to directly transferring a simple bit image, the extra image processing incurred imposes a major reduction in overall printing speed. That is, the image output speed of the printer is now substantially determined, or limited, by the speed with which the processor performs image processing and memory access times which combine to create much slower transfer rates than the xerography unit is capable of handling, resulting in a major reduction in printing capability.
For example, when considering a page printer rated for ten pages a minute, no more than six seconds are allowed for processing image data for each page to be printed. To process all 0.9 megabytes of stored data into an image within this time span, only allows 6.67 microseconds of processing time per byte of data (6 seconds divided by 0.9 megabytes). Such short processing periods represent a processing capacity that may or may not be realizable even with currently available high-speed RISC type processors. In contrast to this processing limitation, the electrostatic charge and photosensitive elements of a laser printer are often capable of easily printing ten or more pages per minute. As a result, under the current state of the art, the processing capability of the image data control unit represents a major stumbling block in improving overall printing speed.
Many page printers are provided with either an internal memory expansion capability or an expansion slot to provide some additional processing capacity. Where an expansion slot is provided, insertion of an "add-on" or expansion "cartridge", here containing font information or a program, expands printer functionality. The addition of pre-formed fonts and font control language to the printer may speed image formation by alleviating the need for some image processing steps. However, even if processing speed is increased using some form of memory expansion, it is not possible to improve the processor performance itself. For example, for a laser printer only supporting one particular PDL, it is common knowledge that PDL interpreter programs are available in the form of integrated circuit cards and add-on cartridges for expanding processing functions to accommodate other page description languages. Such cartridges store programs, or special program routines, in mask ROM form for recall during image processing, and are inserted into the expansion slot of the printer. But the basic printer processor is unchanged and may even run slower implementing these routines.
Expansion cartridge slots have a specific address, or address range or space assigned to them which is detected and read by a printer control unit after power is applied to the printer. If a cartridge containing a PDL interpreter program has been inserted, and, therefore, resides at the appropriate addresses, a pre-selected code is returned to the controller to indicate that the cartridge contains a PDL program. In this situation, control of the printer for image data developing switches to the interpreter program which is read from its address locations inside the cartridge. As a result, the printer is able to interpret data received which is based on use of the particular PDL used by the cartridge program. The use of an interpreter program does not itself increase the processing speed and the overall printing speed may in fact decrease as a result of employing a high level description language with the printer processor.
For this and other reasons, a cartridge equipped with at least a second microprocessor separate from that normally used by the printer has been invented to resolve the problems described above. This cartridge and certain of its features are disclosed in the co-pending U.S. patent applications listed above which are incorporated herein by reference. The disclosed cartridge is able to receive print data from the printer and use its own microprocessor to process and develop image data based on stored PDL interpreters and other program data, and then provide print data back to the printer for forming the desired output image.
However, any advanced microprocessor comprises an electronic circuit having from tens to hundreds of thousands of components or elements, such as transistors, which operate, or switch between operating states, at frequencies of 20 MHz to 40 MHz, or higher. As a consequence, such microprocessors typically generate substantial amounts of heat during operation, increasing the operating temperature of the microprocessor structure, and potentially generating errors or causing physical deterioration and destruction if the heat is not adequately dissipated.
In order to prevent processing errors and destruction of the add-on microprocessor device, the microprocessor housing or casing is typically designed to maintain a maximum temperature of about 80.degree. C. In order to maintain the surface temperature of the casing within tolerances, or colder, it is important to devise a cartridge structure that makes it easy to dissipate heat from the microprocessor or other heat generating components within the cartridge to the surrounding environment.
However, to date, expansion cartridges have not used microprocessors and there has been no need for, nor effort expended to create, a cartridge heat dissipation structure. The heat dissipation problem for add-on cartridges or integrated circuit assemblies is not limited to printers but also extends to other add-on products having microprocessors or other sophisticated components. In general, it is a common problem with add-on electronic devices that are installed in most electronic equipment.
What is needed is a new method and apparatus for dissipating heat generated in add-on circuits especially of the cartridge-type and for minimizing heat accumulation.